As researchers begin trying to build devices and novel materials at the nanoscale (a nanometer is a billionth of a meter, the size of a few atoms), they’re facing a massive challenge. While it’s proving possible, in many cases, to push molecules around to form tiny structures and even functioning devices, efficiently mass-producing anything with nanoscale features is another matter altogether. But what if millions of these nano building blocks did the heavy lifting and assembled themselves into the desired structures-avoiding the use of expensive and elaborate manufacturing instruments?

Self-assembly has become one of the holy grails of nanotechnology, and scientists in numerous labs are working to transform it into an effective nano engineering tool. In some sense self-assembly is nothing new: biology does it all the time. And for decades, scientists have studied “supramolecular” chemistry, learning not only how molecules bind to one another but how large numbers of molecules can team up to form structures; in fact, the concept of self-assembly largely grew out of chemists’ attempts to make molecules that aggregated spontaneously into specific configurations, in the same way biological molecules form complex cell membranes.

But now, with an expanding understanding of how molecules and small particles interact with one another, researchers can begin to predict how such elements might self-assemble into larger, useful structures like the transistors on a semiconductor chip. “Self-assembly provides a very general route to fabricating structures from components too small or too numerous to be handled robotically,” says George Whitesides, a chemist at Harvard University and pioneer in the field.

To better understand how self-assembly works, Whitesides and his coworkers have recently shown that selectively coating the surfaces of microscopic gold plates with a sticky organic film can, under the proper conditions, trigger thousands of such plates to self-assemble into three-dimensional structures. So far, Whitesides’s team has created a relatively large functional electronic circuit using a similar technique. The next step will involve shrinking the circuit to the micrometer scale, creating more complex three-dimensional structures out of silicon. While micrometer-sized electronic components are nothing new-Intel makes them all the time-Whitesides’s experiments could provide valuable clues as to how to better manipulate self-assembly.

Nature itself is also providing scientists with a model of how to create self-assembling electronic devices. Materials scientist Angela Belcher at the University of Texas at Austin sorted through billions of different proteins to find ones that recognize and bind to different types of inorganic materials. For instance, one end of the protein might bind to a specific metal particle and the other end might stick to the surface of a semiconductor such as gallium arsenide. Given the right prompts, the proteins could direct nano-sized particles of inorganic materials to form various structures.

This past spring, Belcher cofounded a company called Semzyme that plans to create a library of these protein-mediated building blocks. They could have any number of technological applications, in making such things as biomedical sensors, high-density magnetic storage disks or microprocessors.

Chemists at labs such as those of Hewlett-Packard, the University of California, Los Angeles, Yale University and Rice University are also attempting to develop self-assembled molecular computers. If they succeed, however, it will take years.

Meanwhile, less ambitiously, other researchers are making rapid strides in using self-assembly to build increasingly complex-and increasingly small-three-dimensional structures that could be compatible with existing devices. For instance, certain features of a disk drive, like the storage medium, could be created using self-assembly, while larger components needed to connect the device to the outside world would be made using conventional techniques. “We hope that self-assembly will be able to inexpensively replace certain stages in the production of materials and devices, where control is needed at the molecular level,” says engineer Christopher Murray of the nanoscale-science division of IBM Research in Yorktown Heights, NY.

If he’s right, nano engineering will get a whole lot easier.